2021-02-28 12:33:03

by Dinghao Liu

Subject: [PATCH] iwlegacy: Add missing check in il4965_commit_rxon

There is one il_set_tx_power() call in this function without
return value check. Print error message and return error code
on failure just like the other il_set_tx_power() call.

Signed-off-by: Dinghao Liu <[email protected]>
---
drivers/net/wireless/intel/iwlegacy/4965.c | 6 +++++-
1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/drivers/net/wireless/intel/iwlegacy/4965.c b/drivers/net/wireless/intel/iwlegacy/4965.c
index 9fa556486511..3235b8be1894 100644
--- a/drivers/net/wireless/intel/iwlegacy/4965.c
+++ b/drivers/net/wireless/intel/iwlegacy/4965.c
@@ -1361,7 +1361,11 @@ il4965_commit_rxon(struct il_priv *il)
* We do not commit tx power settings while channel changing,
* do it now if tx power changed.
*/
- il_set_tx_power(il, il->tx_power_next, false);
+ ret = il_set_tx_power(il, il->tx_power_next, false);
+ if (ret) {
+ IL_ERR("Error sending TX power (%d)\n", ret);
+ return ret;
+ }
return 0;
}

--
2.17.1

2021-03-01 07:29:26

by Stanislaw Gruszka

Subject: Re: [PATCH] iwlegacy: Add missing check in il4965_commit_rxon

On Sun, Feb 28, 2021 at 08:25:22PM +0800, Dinghao Liu wrote:
> There is one il_set_tx_power() call in this function without
> return value check. Print error message and return error code
> on failure just like the other il_set_tx_power() call.

We have few calls for il_set_tx_power(), on some cases we
check return on some not. That correct as setting tx power
can be deferred internally if not possible at the moment.

> Signed-off-by: Dinghao Liu <[email protected]>
> ---
> drivers/net/wireless/intel/iwlegacy/4965.c | 6 +++++-
> 1 file changed, 5 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/net/wireless/intel/iwlegacy/4965.c b/drivers/net/wireless/intel/iwlegacy/4965.c
> index 9fa556486511..3235b8be1894 100644
> --- a/drivers/net/wireless/intel/iwlegacy/4965.c
> +++ b/drivers/net/wireless/intel/iwlegacy/4965.c
> @@ -1361,7 +1361,11 @@ il4965_commit_rxon(struct il_priv *il)
> * We do not commit tx power settings while channel changing,
> * do it now if tx power changed.
> */
> - il_set_tx_power(il, il->tx_power_next, false);
> + ret = il_set_tx_power(il, il->tx_power_next, false);
> + if (ret) {
> + IL_ERR("Error sending TX power (%d)\n", ret);
> + return ret;
> +

This is not good change. We do not check return value of
il_commit_rxon(), except when creating interface. So this change might
broke creating interface, what worked otherwise when il_set_tx_power()
returned error.

Stanislaw

2021-03-01 07:43:45

by Dinghao Liu

Subject: Re: Re: [PATCH] iwlegacy: Add missing check in il4965_commit_rxon

> On Sun, Feb 28, 2021 at 08:25:22PM +0800, Dinghao Liu wrote:
> > There is one il_set_tx_power() call in this function without
> > return value check. Print error message and return error code
> > on failure just like the other il_set_tx_power() call.
>
> We have few calls for il_set_tx_power(), on some cases we
> check return on some not. That correct as setting tx power
> can be deferred internally if not possible at the moment.
>
> > Signed-off-by: Dinghao Liu <[email protected]>
> > ---
> > drivers/net/wireless/intel/iwlegacy/4965.c | 6 +++++-
> > 1 file changed, 5 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/net/wireless/intel/iwlegacy/4965.c b/drivers/net/wireless/intel/iwlegacy/4965.c
> > index 9fa556486511..3235b8be1894 100644
> > --- a/drivers/net/wireless/intel/iwlegacy/4965.c
> > +++ b/drivers/net/wireless/intel/iwlegacy/4965.c
> > @@ -1361,7 +1361,11 @@ il4965_commit_rxon(struct il_priv *il)
> > * We do not commit tx power settings while channel changing,
> > * do it now if tx power changed.
> > */
> > - il_set_tx_power(il, il->tx_power_next, false);
> > + ret = il_set_tx_power(il, il->tx_power_next, false);
> > + if (ret) {
> > + IL_ERR("Error sending TX power (%d)\n", ret);
> > + return ret;
> > +
>
> This is not good change. We do not check return value of
> il_commit_rxon(), except when creating interface. So this change might
> broke creating interface, what worked otherwise when il_set_tx_power()
> returned error.
>

It's clear to me, Thanks for your explanation!

Regards,
Dinghao